We propose to develop and test methods for direct computer interviewing for drug use/abuse and drug related behavior and attitudes. The ultimate aim is to provide a fast, economical way to collect drug risk data, combining advantages of paper and pencil and personal interviews. The reliability and validity of the computer interview will be tested in a series of studies on high school and college-age respondents: 1. one week test-retest computer interview; 2. comparison of computer interview and case records at local treatment center; 3. comparison of computer and personal interview; 4. comparison of refusal, completion and response rates and patterns for computer-interview and paper and pencil data in a stratified sample of 1600 (200 each group) high school students in Madison, Boston, Indianapolis, and Denver. Results will be used to design a final drug risk interview for use in future surveys, with immediate applicability to over 35 computer systems nationwide.